Explore warehouse data on a visual canvas
Search for tables, ask questions in plain language and chain queries and charts on a canvas, with each step's SQL visible.
The use case
Exploratory analysis often means many small queries that are hard to keep track of. BigQuery data canvas lets an analyst find tables with natural-language search, ask questions that become SQL, and chain results into further queries and charts on a single canvas. Gemini writes the SQL and summarizes results, and every node shows its query. This keeps the exploration path visible and shareable. The analyst still reviews each query, checks joins and filters, and moves anything important into a reviewed, scheduled query before others depend on it.
Example workflow
- Start a canvas and search for relevant tables
- Ask a question to create a query node
- Branch into follow-up queries and charts
- Review SQL at each step
- Promote final queries to scheduled jobs
Tools
- BigQuery data canvas
- Gemini in BigQuery
Risks and guardrails
- Generated joins can double-count rows
- Canvas results are ad hoc; do not treat as reporting
- Query costs add up; set limits
Sources
Written by Agents Lookup. Credits and official pages for the tools this use case names:
- Google Cloud docs: BigQuery data canvas: docs.cloud.google.com/bigquery/docs/data-canvas